Hello, Sorry not to answer quickly I have had lots of work. Nothing very new since months....
I just do some windows snoops in 1200dpi and 2400dpi translated into C with usbsnoop2libusb.pl. they are : http://ggastebois.free.fr/lide90_snoop/lide90_1200dpi_216x7.c and http://ggastebois.free.fr/lide90_snoop/lide90_2400dpi_216x7.c I don't see big differences.... > Guillaume Gastebois schrieb: >> > Hello, >> > >> > No progress on 2400dpi. DPIHW is 2400 and dpiset 600 for a 300dpi scan. >> > Don't know where to look yet for that problem ! > > I'd try to get a log from a 2400dpi scan from windows and compare. > >> > about byte nibbles modifying reg_0x79 from 0x3f to 0x3e (very few >> > contrasted image) or 0x40(black image) gives me bad images !!! >> > >> > What is effect of this register on clk3 ?? It's very sensible. > > register 0x77 to 0x79 describe a bitmask, that determines when the clk3 > signal is high or low. For 0x0003f(lsb is register 0x79), this would give: > > time===>> ! next pixel > bit of 0x79 ! 0x78 !0x77! 0x79 > 0 1 2 3 4 5 6 7!0 1 2 3 4 5 6 7!0 1!0 1 2 3 4 5 6... > 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 .... > ___________ ___________ > |_______________________| |____ > > 0x40 is this: > > bit of 0x79 ! 0x78 !0x77! 0x79 > 0 1 2 3 4 5 6 7!0 1 2 3 4 5 6 7!0 1!0 1 2 3 4 5 6... > 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 .... > _ _ > ___________| |_________________________________| |____ > > As you can see, this gives a very short pulse, which will probably not > be recognized by the ccd, while 0x3e just shortens the clock pulse at > the beginning. The datasheet mentions that the scanner does not need to > use the full 18 bits of the bitmask, but uses the first 6, 12 or 18 bits > depending on the scanning mode. 12 seems to be correct for the mode the > backend programs for cis scanners. > > Please try if modifying bitmask for clk4(0x7a to 0x7c iirc) shows any > effects. Also, other values you may want to try on clk3: 0xff03f, > 0xff801f, 0xffc00f. Another thing to try may be to send two clock > pulses, for example 0x001c7. If clk3 is connected to the ccd, this may > give interesting results(like, reducing the resolution by half), but for > the 4-bit latch/D-FlipFlop, only the last clock pulse before the gl841 > samples the data bits is relevant. I tryed lot's of differents values for clk3 (differents pulse width and positions) succesless. Only entropy program on offset files gives me sometimes malformed crosses. > >> > regards, >> > Guillaume > > Regards, > Pierre Regards Guillaume